*** Mutual Uncertainty and Credible Reassurance: Experimental Evidence
*** International Interactions
*** Replication Materials

*** data files needed:
*	Agg_Sender_v2.dta
*   Available at: https://www.dropbox.com/s/dqdpx8w1v4oh2uu/Agg_Sender_v2.dta?dl=0

*	Agg_Receiver.dta
*   Available at: https://www.dropbox.com/s/4yggvlbwglusjjq/Agg_Receiver.dta?dl=0

*** Code run using STATA/SE 15.1


*** Note
* Many of the graphs produced with the code below have very long, messy labels ///
* on the x-axis. I cleaned these up using the "Graph Editor" function in STATA plots.

* To replicate this:
* Click "Graph Editor"
* Double Click "xaxis1" tab on right side 
* Click "Edit or add individual ticks or labels"
* Double click each entry in the "Ticks (labels)" field and key in the new label
* Click "Close"
* Click "OK"






****************************** Sender Results **********************************

*** Load SenderData
* Available at: https://www.dropbox.com/s/dqdpx8w1v4oh2uu/Agg_Sender_v2.dta?dl=0

use local file path "/Agg_Sender_v2.dta"


*** Figure 2a: Honesty Hypothesis - Disaggregated Means (Type A Sender)

ciplot honest if type==0, by(s_round) ylabel(0(.2)1) ytitle(Pr Honest Signal) ///
	xtitle(Pr Receiver = Type A) title("")



* Figure 2b: Honesty Hypothesis - Disaggregated Means (Type B Sender)

ciplot honest if type==1, by(s_round) ylabel(0(.2)1) ytitle(Pr Honest Signal) ///
	xtitle(Pr Receiver = Type A) title("")





*** Figure 3a: Honesty Hypothesis - Pooled Results (Pooled Means)

ciplot honest, by(s_round) ylabel(.5(.1)1) ytitle(Pr Honest Signal) ///
	xtitle(Pr Receiver = Type A) title("")



*** Figure 3b: Honesty Hypothesis - Pooled Results (Fitted Model)

logit honest s_round c.s_round#c.s_round r_uncertainty
* NOTE: This code also produces the results for Model 1 in Table 1

margins, at(s_round=(0(.01)1)) atmeans vsquish post

marginsplot, ylabel(.5(.1)1) recast(line) recastci(rarea) legend(pos(11) ring(0) ///
	cols(1) bmargin(small) lstyle(none)) ytitle(Pr Honest Signal) ///
	xtitle(Pr Receiver = Type A) title("")



*** Figure 4: ProbA vs Sender Uncertainty

twoway scatter s_uncertainty s_round 









***************************** Receiver Results *********************************
 
*** Load Data
* Available at: https://www.dropbox.com/s/4yggvlbwglusjjq/Agg_Receiver.dta?dl=0

use local file path "/Agg_Receiver.dta"



*** Figure 5a: Accuracy Hypothesis (Means)

ciplot correct_action, by(s_uncertainty) ylabel(.6(.1)1) ytitle(Pr Correct Guess) ///
	xtitle(Sender's Uncertainty) title("")
	
	
	
*** Figure 5b: Accuracy Hypothesis (Fitted Model)

logit correct s_uncertainty r_uncertainty
*NOTE: This code also produces the results for Model 2 in Table 1

margins, at(s_uncertainty=(0(.01)1)) atmeans vsquish post
marginsplot, recast(line) recastci(rarea) legend(pos(11) ring(0) cols(1) ///
	bmargin(small) lstyle(none)) ylabel(.6(.1)1) ytitle(Pr Correct Guess) ///
	xtitle(Sender's Uncertainty) title("")
	
	
	
*** Figure 6a: Response Hypothesis (Means)

ciplot correct_action, by(s_uncertainty) ylabel(.6(.1)1) ytitle(Pr Correct Action) ///
	xtitle(Sender's Uncertainty) title("")
	
	
	
*** Figure 6b: Response Hypothesis (Fitted Model)

logit correct_action s_uncertainty r_uncertainty
* NOTE: This code also produces the results for Model 3 in Table 1

margins, at(s_uncertainty=(0(.01)1)) atmeans vsquish post

marginsplot, recast(line) recastci(rarea) legend(pos(11) ring(0) cols(1) ///
	bmargin(small) lstyle(none)) ylabel(.6(.1)1) ytitle(Pr Correct Action) ///
	xtitle(Sender's Uncertainty) title("")
 
 
 
 *** Figure 7a: Confidence Hypothesis: Pooled Means (Unincentivized Confidence)
 
 ciplot confidence, by(s_uncertainty) ylabel(71(2)83) ///
	ytitle(Receiver Confidence (Unincentivized)) xtitle(Sender's Uncertainty) title("")
	

	
 *** Figure 7b: Confidence Hypothesis: Pooled Means (Incentivized Confidence)

 ciplot inc_conf, by(s_uncertainty) ylabel(2.5(.5)4.5) ///
	ytitle(Receiver Confidence (Incentivized)) xtitle(Sender's Uncertainty) ///
	title("") 
	
	
	
*** Figure 8a: Confidence Hypothesis: Fitted Results (Unincentivized)

reg confidence s_uncertainty r_uncertainty
* NOTE: This code also produces the results for Model 4 in Table 1
margins, at(s_uncertainty=(0(.01)1)) atmeans vsquish post
eststo c1



reg confidence s_uncertainty r_uncertainty
margins, at(r_uncertainty=(0(.01)1)) atmeans vsquish post
eststo c2

coefplot c1 c2, vertical recast(line) legend(pos(1) ring(0) cols(1) ///
	bmargin(small) lstyle(none)) ytitle(Unincentivized Confidence) xtitle(Uncertainty) ///
	ylabel(50(10)100) 

	
	
	
*** Figure 8b: Confidence Hypothesis: Fitted Results (Incentivized)

ologit inc_conf s_uncertainty r_uncertainty
* NOTE: This code also produces the results for Model 5 in Table 1
margins, predict(outcome(5)) at(s_uncertainty=(0(.01)1)) atmeans vsquish post
eststo c3


ologit inc_conf s_uncertainty r_uncertainty
margins, predict(outcome(5)) at(r_uncertainty=(0(.01)1)) atmeans vsquish post
eststo c4

coefplot c3 c4, vertical recast(line) legend(pos(1) ring(0) cols(1) ///
	bmargin(small) lstyle(none)) ytitle(Pr Incentivized Confidence = 5) xtitle(Uncertainty) ///
	ylabel(0(.2)1) 



	
	
	
	
	
	
	
********************************* Appendix *************************************

*** Sender Results 

*** Load Sender Data
* Available at: https://www.dropbox.com/s/dqdpx8w1v4oh2uu/Agg_Sender_v2.dta?dl=0

use local file path "/Agg_Sender_v2.dta"


*** Table 3: Honesty Hypothesis

* Model 1
reg honest s_round c.s_round#c.s_round r_uncertainty
estat ic

* Model 2
reg honest s_round c.s_round#c.s_round r_uncertainty age gpa gender experience period
estat ic

* Model 3
reg honest s_uncertainty r_uncertainty
estat ic

* Model 4
reg honest s_uncertainty r_uncertainty age gpa gender experience period
estat ic

* Model 5
logit honest s_round c.s_round#c.s_round r_uncertainty
estat ic

* Model 6
logit honest s_round c.s_round#c.s_round r_uncertainty age gpa gender experience period
estat ic

* Model 7
logit honest s_uncertainty r_uncertainty
estat ic

* Model 8
logit honest s_uncertainty r_uncertainty age gpa gender experience period
estat ic


*** Figure 2: Honesty Hypothesis: Predicted Probabilities


*** Figure 2a: OLS - Model 4
reg honest s_uncertainty r_uncertainty age gpa gender experience period

margins, at(s_uncertainty=(0(.01)1)) atmeans vsquish post

marginsplot, ylabel(.6(.1)1) recast(line) recastci(rarea) legend(pos(11) ring(0) ///
	cols(1) bmargin(small) lstyle(none)) ytitle(Pr Honest Signal) ///
	xtitle(Sender's Uncertainty) title("")
	
	

*** Figure 2b: Logit - Model 8
logit honest s_uncertainty r_uncertainty age gpa gender experience period

margins, at(s_uncertainty=(0(.01)1)) atmeans vsquish post

marginsplot, ylabel(.6(.1)1) recast(line) recastci(rarea) legend(pos(11) ring(0) ///
	cols(1) bmargin(small) lstyle(none)) ytitle(Pr Honest Signal) ///
	xtitle(Sender's Uncertainty) title("")



*** Figure 2c: OLS - Model 2
reg honest s_round c.s_round#c.s_round r_uncertainty age gpa gender experience period

margins, at(s_round=(0(.01)1)) atmeans vsquish post

marginsplot, ylabel(.6(.1)1) recast(line) recastci(rarea) legend(pos(11) ring(0) ///
	cols(1) bmargin(small) lstyle(none)) ytitle(Pr Honest Signal) ///
	xtitle(Sender's Uncertainty) title("")
	
	
	
*** Figure 2d: Logit - Model 6
reg honest s_round c.s_round#c.s_round r_uncertainty age gpa gender experience period

margins, at(s_round=(0(.01)1)) atmeans vsquish post

marginsplot, ylabel(.6(.1)1) recast(line) recastci(rarea) legend(pos(11) ring(0) ///
	cols(1) bmargin(small) lstyle(none)) ytitle(Pr Honest Signal) ///
	xtitle(Sender's Uncertainty) title("")
	
	
clear
	
	
	
	
*** Receiver Results

*** Load Receiver Data
* Available at: https://www.dropbox.com/s/4yggvlbwglusjjq/Agg_Receiver.dta?dl=0

use local file path "/Agg_Receiver.dta"


*** Table 4: Accuracy Hypothesis

* Model 9
reg correct s_uncertainty r_uncertainty
estat ic

* Model 10
reg correct s_uncertainty r_uncertainty age gpa gender experience period
estat ic

* Model 11
logit correct s_uncertainty r_uncertainty
estat ic

* Model 12
logit correct s_uncertainty r_uncertainty age gpa gender experience period
estat ic



*** Figure 3: Accuracy Hypothesis - Predicted Probabilities

*** Figure 3a: OLS

reg correct s_uncertainty r_uncertainty age gpa gender experience period

margins, at(s_uncertainty=(0(.01)1)) atmeans vsquish post

marginsplot, ylabel(.7(.05).9) recast(line) recastci(rarea) legend(pos(11) ring(0) ///
	cols(1) bmargin(small) lstyle(none)) ytitle(Pr Correct Guess) ///
	xtitle(Sender's Uncertainty) title("")
	
	
*** Figure 3a: Logit

logit correct s_uncertainty r_uncertainty age gpa gender experience period

margins, at(s_uncertainty=(0(.01)1)) atmeans vsquish post

marginsplot, ylabel(.7(.05).9) recast(line) recastci(rarea) legend(pos(11) ring(0) ///
	cols(1) bmargin(small) lstyle(none)) ytitle(Pr Correct Guess) ///
	xtitle(Sender's Uncertainty) title("")
	
	
	
	
	
*** Table 5: Response Hypothesis


* Model 13
reg correct_action s_uncertainty r_uncertainty
estat ic

* Model 14
reg correct_action s_uncertainty r_uncertainty age gpa gender experience period
estat ic

* Model 15
logit correct_action s_uncertainty r_uncertainty
estat ic

* Model 16
logit correct_action s_uncertainty r_uncertainty age gpa gender experience period
estat ic



*** Figure 4a: OLS

reg correct_action s_uncertainty r_uncertainty age gpa gender experience period

margins, at(s_uncertainty=(0(.01)1)) atmeans vsquish post

marginsplot, ylabel(.6(.1)1) recast(line) recastci(rarea) legend(pos(11) ring(0) ///
	cols(1) bmargin(small) lstyle(none)) ytitle(Pr Correct Action) ///
	xtitle(Sender's Uncertainty) title("")
	
	
	
*** Figure 4b: Logit

logit correct_action s_uncertainty r_uncertainty age gpa gender experience period

margins, at(s_uncertainty=(0(.01)1)) atmeans vsquish post

marginsplot, ylabel(.6(.1)1) recast(line) recastci(rarea) legend(pos(11) ring(0) ///
	cols(1) bmargin(small) lstyle(none)) ytitle(Pr Correct Action) ///
	xtitle(Sender's Uncertainty) title("")
	
	
	
	
	
	
*** Table 6: Response Hypothesis


* Model 17
logit condition s_uncertainty r_uncertainty 
estat ic

* Model 18
logit condition s_uncertainty r_uncertainty age gpa gender experience period 
estat ic

* Model 19
reg condition s_uncertainty r_uncertainty 
estat ic

* Model 20
reg condition s_uncertainty r_uncertainty age gpa gender experience period 
estat ic




*** Figure 5a: Means

ciplot condition, by(s_uncertainty) ylabel(.6(.1)1) ytitle(Pr Reciprocate) ///
	xtitle(Sender's Uncertainty)



*** Figure 5b: OLS

reg condition s_uncertainty r_uncertainty age gpa gender experience period 

margins, at(s_uncertainty=(0(.01)1)) atmeans vsquish post

marginsplot, ylabel(.6(.1)1) recast(line) recastci(rarea) legend(pos(11) ring(0) ///
	cols(1) bmargin(small) lstyle(none)) ytitle(Pr Correct Action) ///
	xtitle(Sender's Uncertainty) title("")
	
	
	
*** Figure 5c: Logit

logit condition s_uncertainty r_uncertainty age gpa gender experience period 

margins, at(s_uncertainty=(0(.01)1)) atmeans vsquish post

marginsplot, ylabel(.5(.1)1) recast(line) recastci(rarea) legend(pos(11) ring(0) ///
	cols(1) bmargin(small) lstyle(none)) ytitle(Pr Correct Action) ///
	xtitle(Sender's Uncertainty) title("")
	
	
	
	
	
*** Confidence Hypothesis

*** Table 7: Sender Uncertainty and Receiver Guess Confidence

* Model 21
reg confidence s_uncertainty r_uncertainty
estat ic

* Model 22
reg confidence s_uncertainty r_uncertainty age gpa gender experience period
estat ic

* Model 23
reg inc_conf s_uncertainty r_uncertainty
estat ic

* Model 24
reg inc_conf s_uncertainty r_uncertainty age gpa gender experience period
estat ic

* Model 25
ologit inc_conf s_uncertainty r_uncertainty
estat ic

* Model 26
ologit inc_conf s_uncertainty r_uncertainty age gpa gender experience period
estat ic




*** Figure 6a: Confidence Hypothesis - Predicted Probabilities (Unincentivized)

reg confidence s_uncertainty r_uncertainty age gpa gender experience period

margins, at(s_uncertainty=(0(.01)1)) atmeans vsquish post

marginsplot, ylabel(72(2)80) recast(line) recastci(rarea) legend(pos(11) ring(0) ///
	cols(1) bmargin(small) lstyle(none)) ytitle(Receiver's Confidence (Unincentivized)) ///
	xtitle(Sender's Uncertainty) title("")
	
	
	
	
*** Figure 6a: Confidence Hypothesis - Predicted Probabilities (Incentivized)

reg inc_conf s_uncertainty r_uncertainty age gpa gender experience period

margins, at(s_uncertainty=(0(.01)1)) atmeans vsquish post

marginsplot, ylabel(3.2(.1)4) recast(line) recastci(rarea) legend(pos(11) ring(0) ///
	cols(1) bmargin(small) lstyle(none)) ytitle(Receiver's Confidence (Incentivized)) ///
	xtitle(Sender's Uncertainty) title("")

	
clear
	
	
	
	
*** Table 8: Learning Effects


*** Load Sender Data
* Available at: https://www.dropbox.com/s/dqdpx8w1v4oh2uu/Agg_Sender_v2.dta?dl=0

use local file path "/Agg_Sender_v2.dta"


* Model 27
logit honest c.s_uncertainty#c.period s_uncertainty r_uncertainty period


*** Load Receiver Data
* Available at: https://www.dropbox.com/s/4yggvlbwglusjjq/Agg_Receiver.dta?dl=0

use local file path "/Agg_Receiver.dta"


* Model 28
logit correct c.s_uncertainty#c.period s_uncertainty r_uncertainty period

* Model 29
logit correct_action c.s_uncertainty#c.period s_uncertainty r_uncertainty period

* Model 30
logit condition c.s_uncertainty#c.period s_uncertainty r_uncertainty period

* Model 31
reg confidence c.s_uncertainty#c.period s_uncertainty r_uncertainty period

* Model 32
ologit inc_conf c.s_uncertainty#c.period s_uncertainty r_uncertainty period

clear




*** Figure 7: Learning Effects - Predicted Probabilities


*** Load Sender Data
* Available at: https://www.dropbox.com/s/dqdpx8w1v4oh2uu/Agg_Sender_v2.dta?dl=0

use local file path "/Agg_Sender_v2.dta"


*** Figure 7a: Sender's Honest Signal

logit honest c.s_uncertainty#c.period s_uncertainty r_uncertainty period

margins, at(s_uncertainty=(0(.01)1) period=(3 25)) atmeans vsquish post

marginsplot, recast(line) recastci(rarea) legend(pos(11) ring(0) cols(1) bmargin(small) ///
	lstyle(none)) ytitle(Pr Correct Guess) xtitle(Sender's Uncertainty) title("") ///
	ylabel(.5(.1)1)

clear



*** Load Receiver Data
* Available at: https://www.dropbox.com/s/4yggvlbwglusjjq/Agg_Receiver.dta?dl=0

use local file path "/Agg_Receiver.dta"


*** Figure 7b: Receiver's Correct Guess

logit correct c.s_uncertainty#c.period s_uncertainty r_uncertainty period

margins, at(s_uncertainty=(0(.01)1) period=(3 25)) atmeans vsquish post

marginsplot, recast(line) recastci(rarea) legend(pos(11) ring(0) cols(1) bmargin(small) ///
	lstyle(none)) ytitle(Pr Correct Guess) xtitle(Sender's Uncertainty) title("") ///
	ylabel(.5(.1)1)

	
	
*** Figure 7c: Receiver's Correct action
logit correct_action c.s_uncertainty#c.period s_uncertainty r_uncertainty period

margins, at(s_uncertainty=(0(.01)1) period=(3 25)) atmeans vsquish post

marginsplot, recast(line) recastci(rarea) legend(pos(11) ring(0) cols(1) bmargin(small) ///
	lstyle(none)) ytitle(Pr Correct Action) xtitle(Sender's Uncertainty) title("") ///
	ylabel(.5(.1)1)
	
	
	
*** Figure 7d: Receiver's Conditional action
logit condition c.s_uncertainty#c.period s_uncertainty r_uncertainty period

margins, at(s_uncertainty=(0(.01)1) period=(3 25)) atmeans vsquish post

marginsplot, recast(line) recastci(rarea) legend(pos(11) ring(0) cols(1) bmargin(small) ///
	lstyle(none)) ytitle(Pr Correct Action) xtitle(Sender's Uncertainty) title("") ///
	ylabel(.6(.1)1)

	
	
*** Figure 7e: Receiver's Unincentivized Confidence
reg confidence c.s_uncertainty#c.period s_uncertainty r_uncertainty period

margins, at(s_uncertainty=(0(.01)1) period=(3 25)) atmeans vsquish post

marginsplot, recast(line) recastci(rarea) legend(pos(11) ring(0) cols(1) bmargin(small) ///
	lstyle(none)) ytitle(Pr Correct Action) xtitle(Sender's Uncertainty) title("") ///
	ylabel(71(2)83)

	
	
*** Figure 7f: Receiver's Incentivized Confidence
ologit inc_conf c.s_uncertainty#c.period s_uncertainty r_uncertainty period

margins, predict(outcome(5)) at(s_uncertainty=(0(.01)1) period=(3 25)) atmeans vsquish post

marginsplot, recast(line) recastci(rarea) legend(pos(11) ring(0) cols(1) bmargin(small) ///
	lstyle(none)) ytitle(Pr Correct Action) xtitle(Sender's Uncertainty) title("") ///
	ylabel(.3(.1).8)




	
	
	






